What it is
Legal personhood is a status the law assigns to an entity so it can hold rights and duties, own property, enter contracts, and sue or be sued in its own name. It has nothing to do with consciousness or biology. The law has always distinguished between “natural persons” (human beings) and “legal” or “juridical persons” — artificial constructs the law treats as if they were persons for specific purposes. Corporations are the best-known example: a company can own assets, be sued for negligence, and enter contracts entirely separately from the people who work there or hold its shares.
The reason legal systems extend personhood to non-human entities is practical, not philosophical. Bundling a corporation’s liabilities and assets into a single legal “person” lets courts assign responsibility cleanly, lets investors limit their exposure, and lets contracts function without naming every individual involved. States, government agencies, and in a handful of jurisdictions natural features such as rivers have also been granted forms of legal personhood for similarly functional reasons — to make them legally addressable, not because the law considers them sentient.
That functional logic is exactly what makes AI systems a hard case. An AI agent is neither a natural person nor an obviously bounded entity like a corporation, yet it can take actions — executing trades, drafting contracts, interacting with other systems — that produce real-world consequences. The question legal scholars and regulators have been circling for roughly a decade is whether some form of personhood, or a bespoke equivalent, is the right tool for assigning responsibility when an AI system causes harm — or whether it is a tool that produces the wrong incentives entirely.
Why it matters for AI governance and narratives
The personhood question sits directly on the fault line between two framings of AI harm: one in which responsibility flows to the humans and companies that build and deploy a system, and one in which the system itself becomes a legally addressable actor. The European Parliament’s 2017 resolution on civil law rules for robotics floated “electronic personality” for sufficiently autonomous machines — a status that would have let a robot or algorithm bear liability in its own right. Hundreds of AI, robotics, law, and ethics experts signed an open letter opposing the idea in 2018, arguing it risked becoming a liability shield: if the machine is the legal person, the humans who built, sold, and profited from it become harder to hold to account. The EU abandoned the proposal, and the AI Act that followed instead designates AI systems as regulated products, placing obligations squarely on developers and deployers rather than on the technology itself.
That 2017–2019 fight is the direct backdrop to Pope Leo XIV’s remarks this week. His argument — that granting AI agents rights or personhood “risks obscuring human responsibility and leaving those who suffer harm to confront a faceless entity” — echoes the liability-shield concern from the EU debate almost exactly, from a different institutional vantage point. For the observatory’s purposes, the personhood question is a useful marker of where an actor sits on AI accountability more broadly: proposals to grant AI systems personhood or rights tend to originate from discussions about autonomy and moral status, while proposals to deny it tend to originate from discussions about liability and control. Watching who raises the personhood question, and why, is a reasonably reliable signal of which of those two conversations they are actually having.
Key facts and dates
The European Parliament adopted its Civil Law Rules on Robotics resolution on 16 February 2017, which included the electronic-personality proposal; the 2018 open letter against it was signed by legal scholars, ethicists, and robotics and AI researchers; the European Parliament’s subsequent 2019 resolution on AI industrial policy dropped the personhood idea, and the EU AI Act later codified AI as a regulated entity rather than a legal person. No major jurisdiction currently grants AI systems legal personhood, though the question continues to surface in academic and policy literature as AI agents take on more autonomous, consequential tasks (contracting, financial transactions, agentic workflows).
Pope Leo XIV’s comments came on 30 October 2026, in an address to the Vatican meeting “The Future of AI: A Dialogue Between the Vatican and the Scientific World,” organized jointly by the Pontifical Academy for Life and the Future of Life Institute. He stated plainly that “machines are tools, not persons,” and tied the rejection of AI personhood to a broader call — which he has described as AI “disarmament” — for keeping AI systems under human oversight and accountability rather than treating them as autonomous legal actors.
Where to learn more
- European Parliament resolution P8_TA(2017)0051, Civil Law Rules on Robotics — the primary legislative text that proposed, and the EU later abandoned, “electronic personality” for autonomous machines
- Britannica, “Legal personhood” — reference explanation of the general legal concept and its extension beyond natural persons
- Corporate personhood (Wikipedia) — background on the oldest and most established form of non-human legal personhood, the model most AI-personhood proposals borrow from or react against
- Pope warns AI agents are ‘tools, not persons’ — reporting on Pope Leo XIV’s 30 October 2026 Vatican remarks, with direct quotes